I built the A75 almost 15 years ago and it has operated flawlessly without any problems up until a couple weeks ago. After running a few tests I discovered a bad 10 ohm resistor in the +Reg supply. Before reassembling everything I thought I would recalibrate the front end..adjust the DC offset and check the bias and this is where I ran into a problem. The lowest I can get the DC offset on 1 channel is -1.5VDC, which is way too high. Measurements taken across the resisitors and zeners as per the instructions for testing the front end are as they should be (1V, 5V, etc..). I pulled the diff pair input transistors and checked their VGS for matching within 30mv and all were fine. Before I start pulling any more parts and replacing them I thought I would check with the group here to see if anyone has any suggestions. This amp is outstanding and there's no way I am going to retire it.
Any help would be greatly appreciated.
Any help would be greatly appreciated.
What did you have at the inputs? Remember this amp is DC coupled and will amplify any DC at the input.
Please list the voltages you tested. R1, 6, 9, 10 all ~5V? If they are all in the neighborhood of 5V and symmetrical rail to rail, the amp front end is working, it's just giving you adjustment fits. Have you disconnected the output stage drive while adjusting?
How close are the voltages you get on R7 and R8? The amp is super sensitive to imbalances there. Was there a breeze? Forced air heat? My A75s are incredibly sensitive to any air motion. It also seem s like the mere act of lifting the screwdriver from P1 and P2 significantly changes the R7 R8 voltage. Remember that you are trying to balance opposing constant current sources, so if one is up too high, the voltage will pull that direction. You might try cranking them both back to .5V and working your way up to .9V again keeping the offset near zero. Just one of those things.
Are you operating unbalanced? Double check the the inverting input is grounded.
Please list the voltages you tested. R1, 6, 9, 10 all ~5V? If they are all in the neighborhood of 5V and symmetrical rail to rail, the amp front end is working, it's just giving you adjustment fits. Have you disconnected the output stage drive while adjusting?
How close are the voltages you get on R7 and R8? The amp is super sensitive to imbalances there. Was there a breeze? Forced air heat? My A75s are incredibly sensitive to any air motion. It also seem s like the mere act of lifting the screwdriver from P1 and P2 significantly changes the R7 R8 voltage. Remember that you are trying to balance opposing constant current sources, so if one is up too high, the voltage will pull that direction. You might try cranking them both back to .5V and working your way up to .9V again keeping the offset near zero. Just one of those things.
Are you operating unbalanced? Double check the the inverting input is grounded.
Here is the test setup: Running unbalanced with no signal source attached at this point in the test.
Measured voltages across R7,8 = .9VDC, R9,10 approx 5VDC, Z1,2=9VDC.
+ and - Reg VDC at 51.
Output stage, +DRv and -DRV and the ground to it are not connected.
The Bal, -In, and ground next to it are tied to grnd.
I'm going from memory but the VDC across +DRV and -DRV is lower than the 7V it should be. I wanna say about 3.5V (forgot the polarity), which
suggests maybe one of the outputs is out (Q3 or Q6?). I replaced Q3 and Q9 last night thinking that since the +Reg supply failed the front end problem may also be with that side of the circuit. No change.
Is it possible I have a ground issue? the ground connection is mechanical and oer time may have been compromised due to corrosion.
I've adjusted P1 and P2 all the way from .5V to over 1V and the DC reading at the output won't budge. Adjustment senssitivity is extremely low, not how I recall when making this same adjustment the first time.
I would also mention that when I check the output with the scope probe (either channel) the DC does drop...to maybe 1V on the problem channel.
Measured voltages across R7,8 = .9VDC, R9,10 approx 5VDC, Z1,2=9VDC.
+ and - Reg VDC at 51.
Output stage, +DRv and -DRV and the ground to it are not connected.
The Bal, -In, and ground next to it are tied to grnd.
I'm going from memory but the VDC across +DRV and -DRV is lower than the 7V it should be. I wanna say about 3.5V (forgot the polarity), which
suggests maybe one of the outputs is out (Q3 or Q6?). I replaced Q3 and Q9 last night thinking that since the +Reg supply failed the front end problem may also be with that side of the circuit. No change.
Is it possible I have a ground issue? the ground connection is mechanical and oer time may have been compromised due to corrosion.
I've adjusted P1 and P2 all the way from .5V to over 1V and the DC reading at the output won't budge. Adjustment senssitivity is extremely low, not how I recall when making this same adjustment the first time.
I would also mention that when I check the output with the scope probe (either channel) the DC does drop...to maybe 1V on the problem channel.
Im going from memory here on the test readings. Z1,2=9V, R7,8 approx .9V, R9,10=5V.
running unbalanced with -In to grnd, no input connected to +In. no output stage connected. The DC sensitivity and R7,8 V sensitivity when adjusting P1,2 is very low. Big adjustment with little or no DC V change.
I did swap out Q3 and Q9 last night thinking that because the +Reg is what failed the frontend + circuit may have failed somewhere. No change when I retested.
Now you've got me thinking I should check the ground connection since they're mechanically connected at the the BIG supply cap ground plate. 15 years.. could have some corrosion.
running unbalanced with -In to grnd, no input connected to +In. no output stage connected. The DC sensitivity and R7,8 V sensitivity when adjusting P1,2 is very low. Big adjustment with little or no DC V change.
I did swap out Q3 and Q9 last night thinking that because the +Reg is what failed the frontend + circuit may have failed somewhere. No change when I retested.
Now you've got me thinking I should check the ground connection since they're mechanically connected at the the BIG supply cap ground plate. 15 years.. could have some corrosion.
Sounds like the basic operation is correct, currents flowing where they are supposed to.
P1 and P2 being insensitive sounds like you may have lost the wiper contact. 15 years without moving, not terribly unlikely. Do you have any spares? Others I have discussed their experience setting up this amp with came to the same conclusion - no more single turn pots, ever. If the Pot wipers aren't adjusting, you're relying on the balance of the two differentials and Q3/Q6 Vgs to set your VAS current.
Is the problem just with one channel? Old Colony/AudioXpress boards or some other?
It doesn't take much current for the Vgs multiplier to work, if you have 1V across R7/8 there's plenty. You might have knocked P3's adjustment. If the drive voltage doesn't change with P3 adjustment, try replacing Q11. Not sure why that would mess up offset if bad, but you never know.
P1 and P2 being insensitive sounds like you may have lost the wiper contact. 15 years without moving, not terribly unlikely. Do you have any spares? Others I have discussed their experience setting up this amp with came to the same conclusion - no more single turn pots, ever. If the Pot wipers aren't adjusting, you're relying on the balance of the two differentials and Q3/Q6 Vgs to set your VAS current.
Is the problem just with one channel? Old Colony/AudioXpress boards or some other?
It doesn't take much current for the Vgs multiplier to work, if you have 1V across R7/8 there's plenty. You might have knocked P3's adjustment. If the drive voltage doesn't change with P3 adjustment, try replacing Q11. Not sure why that would mess up offset if bad, but you never know.
Last edited:
Connect the grounds - On the AudioXpress boards the one right next to the output is the main ground. The ground connection next to the inputs is for the output Zobel, but won't hurt if connected. Your -input is ground connection isn't connected to the board ground. The board ground should connect to the star.
Purchased this board from Old Colony(?) It has the same layout as the one in the A75 articles. I had a hunch that those variable pots might have gone bad and you just confirmed my suspicion. Let me try replacing them and get back to you with the results. I know I have some spares in the box and if I have enough and this is the problem they'll all get replaced. Sounds like multi-turns are the way to go but will probably have to order some. I have the ground between the -In and Out tied to ground. The GND between the -DRV and OUT is not connected.
The problem is just with the one channel, albeit the adjustment on the other channel is also a big insensitive. Also, I reset the P3's the full resistance when I started with the intent to readjust them. Just haven't gotten that far along yet since it requires connecting the output stage.
I'll try out your suggestions in the next day or 2 and let you know how things work out. I really appreciate your helping me out.
The problem is just with the one channel, albeit the adjustment on the other channel is also a big insensitive. Also, I reset the P3's the full resistance when I started with the intent to readjust them. Just haven't gotten that far along yet since it requires connecting the output stage.
I'll try out your suggestions in the next day or 2 and let you know how things work out. I really appreciate your helping me out.
ok. I think we're back in the game. I spent a couple hours last night dialing in those pesky trim pots. It doesn't appear they were the problem so much as my ground connection. After disconnecting my front end ground wire connections from the big mains supply cap plate and polishing things up I was able to adjust the DC offset and bias. It took me a solid hour to get things dialed in and I can't say I'm overly please with the results. As you put it, a slight breeze across those pots and everything changes. Looks like I need to hunt down some multi-turn drop in replacements for those 5k pots. I checked the Digikey site and didn't see anything that would just drop in for a 5k. If anyone has a supplier and / or part# for a drop in replacement I'd appreciate hearing from you.
Thanks again for your help Bob.
Thanks again for your help Bob.
Glad you're back listening. If you find a top adjust multi turn pot let me know. All the ones I have found are side adjust.
If you got it in an hour you got it a lot faster than me. I just tweaked mine back up to full bias. You'd think it was a simple matter of turning up P3, but the extra heat gave me a bunch of offset. Guess the circulation inside the amp is different to each side of the board. Pulling the top off the enclosure shifts offset ~.2V (more the longer the top was off), so it felt really odd dialing in that much offset. Put the top on, let it heat up 30-40 minutes and see where it drifted. It took 6-8 cycles to get the bias where I wanted it with acceptable offset. The good news is that the amps sound much better at high bias. Previous speakers didn't resolve the difference, so I had left the bias down.
If you got it in an hour you got it a lot faster than me. I just tweaked mine back up to full bias. You'd think it was a simple matter of turning up P3, but the extra heat gave me a bunch of offset. Guess the circulation inside the amp is different to each side of the board. Pulling the top off the enclosure shifts offset ~.2V (more the longer the top was off), so it felt really odd dialing in that much offset. Put the top on, let it heat up 30-40 minutes and see where it drifted. It took 6-8 cycles to get the bias where I wanted it with acceptable offset. The good news is that the amps sound much better at high bias. Previous speakers didn't resolve the difference, so I had left the bias down.
if you're looking for multi-turn 5k trimpots I would try Mouser. I compared to pinout to the square singles installed now and they look to be the same. Here is the link. I just ordered 6. They're not 25 turn but better than the single installed now.
3339P-1-472LF Bourns Trimmer Resistors - Multi Turn
3339P-1-472LF Bourns Trimmer Resistors - Multi Turn
There is about 3mV rms on the outputs with the inputs shorted and its audible when you get your ear up to the woofer. I will probe around the grounds and see if I can find some improvement.
By the way. I have never had much problem dialing in the offset and bias and temperature changes with respect to output offset track well, which is one of the benefits of dual differentials.
By the way. I have never had much problem dialing in the offset and bias and temperature changes with respect to output offset track well, which is one of the benefits of dual differentials.
Is the hum 60 Hz or 120? If you don't have a scope, to me 60 Hz sounds like a nice hum, 120 a bit buzzy. Also, 60 Hz is what you get when you touch the open input.
Did you try to duplicate the chassis in the article? I suspect that inside there is a shielding shelf above the transformer and PSU on which the front end board resides centered between the heat sinks. All in all, get as much space as you can between the mains circuitry and the front end board. Do you have decoupling caps added at the power connections to the front end board?
Did you try to duplicate the chassis in the article? I suspect that inside there is a shielding shelf above the transformer and PSU on which the front end board resides centered between the heat sinks. All in all, get as much space as you can between the mains circuitry and the front end board. Do you have decoupling caps added at the power connections to the front end board?
Ok...I'm almost there but am still struggling with the DC Offset. With .9V across both R7 and R8 i get -1v offset and 3.5V to ground from +DRV and -DRV. To get the offset down to less than 50mv I have to take the bias down across R8 to almost 0v. Nelson stated it wasn't necessary to match the TO220's but only make sure the VGS is between 4-4.5V. The best I could do was 4.5V on the N's and 3.5V on the P's. the inputs are matched to with 30mv. Is it possible this VGS difference between the P/N is why I'm struggling with the offset? And, if so how would I go about compensating for this difference? I'v spent half the day trying to figure this out. Any suggestions would be really appreciated. I would also mention that when adjusting P1/2, adjusting one pot also changes the V across the R7/8 on the other side? Is that normal?
Yes, the two are related. Look at the schematic, you'll see that the current in R7 ends up in R8. Yep, it's touchy as all get out because what you are doing is balancing a current source and a current sink.
Ignore the voltage across the DRV terminals for now. Just set it at minimum.
You might want to turn the back down the pots and try bringing the the R7/R8 voltages up together keeping the offset nulled. It can be a pain, but the only way I could get it reasonably close was to try from zero a couple times.
Once you get your offset nulled at full R7/R8 voltage, then reconnect the output stage and bring up the bias slowly.
Ignore the voltage across the DRV terminals for now. Just set it at minimum.
You might want to turn the back down the pots and try bringing the the R7/R8 voltages up together keeping the offset nulled. It can be a pain, but the only way I could get it reasonably close was to try from zero a couple times.
Once you get your offset nulled at full R7/R8 voltage, then reconnect the output stage and bring up the bias slowly.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.